One agent, every channel

Callers agents are not limited to the phone. The same building blocks (scripts, variables, actions, success criteria, webhooks, and contact memory) work across voice, web calls, SMS, WhatsApp, and email. You learn one editor and one results model, and your contacts can move between channels without starting over.

​

Channels at a glance

You choose the channel when you create an agent. Web calls are a setting on a Voice agent, so one agent can take phone calls and web calls with the same script.

​

What each channel supports

¹ Email agents can also send a human-written template instead of AI-written content. ² Knowledge base retrieval isn’t available for text agents at this time. Put the facts a text agent needs in its script, scenarios, or instructions. ³ Not verified for web calls. Contact support@callers.ai if you need memory on web calls. Voice-only features include voices and emotions, patience and interruption settings, keypad input, call transfers, voicemail handling, and recording. Email adds mailbox rotation, send limits, and an optional human approval queue.

​

Bridge text and voice with pre-call chat

Pre-call chat connects a text channel to an outbound voice agent. Before dialing, the agent sends an SMS or WhatsApp message asking whether the contact is open to a call, then acts on the reply. Contacts who agree expect the call and answer more often. Contacts who decline aren’t called. See Pre-call chat.

​

Contact memory across channels

With contact memory on, the agent saves useful facts from each conversation, such as preferences, commitments, or reasons for declining. The next conversation starts with that context, so the agent doesn’t ask the same questions twice. Memories belong to the contact. Across agents in your organization, they are matched by phone number, so a contact who texted your SMS agent last week can be greeted with that context when your voice agent calls. You can turn off memory per agent, or stop an agent from using memories created by other agents. See Contact memory.

​

One results model

Every channel produces the same kinds of results:

• Outcomes from your success criteria, such as goal reached, not interested, or a custom outcome like “Follow-Up Scheduled.”
• Summaries and key questions from post-conversation analysis.
• Webhooks for each channel’s events, so one endpoint can receive call, chat, and email results. See Webhook events.
• Post-conversation actions that update your CRM or trigger the next step, based on the outcome.

This means you can route a contact from one channel to another with an outcome and an action. For example, send an SMS follow-up when a call ends with “Call me later,” or have an email agent pick up contacts a voice agent couldn’t reach.

​

Build without code, or with the API

Everything on this page can be set up in the dashboard, with help from Studio Chat if you prefer to describe what you want. Developers can create agents, add contacts, start and stop agents, and receive results through the API and webhooks. Web calls can be embedded with the widget or the Web SDK.

​

Related